Recently, a demand to develop a technology of fusing information and contents providers, expansion of consumers, and various information providing media is increased. An example of a representative fusion technology may include a fusion technology of wired and wireless broadcasting and communication technologies, an integrated information web service technology, and a user interface technology development convenient for various user layers. In the detailed contents of the technology, a development of an information market technology is expected so as to meet the taste and intention of the user of the information contents and suggest and provide the information contents by a convenient and various method by automatically analyzing the quality and meaning analysis of contents and information provided or to be provided by a computer.
For example, Korean Patent Application Laid-Open No. 10-2006-0043333 (SYSTEM AND METHOD FOR DETERMINING INTENTION OF DATA AND RESPONDING TO DATA BASED ON INTENTION) discloses a system facilitating data handling. The system is a system including a component receiving data and a data manager that determines the intention of the data, re-systematizes at least a subset of data based on the intention and automatically provides the re-systematized data to a user. The data manager is a system that analyzes the data, sorts the data into at least a set having relevant features, extracts features from at least one of the sets of data to form the subset of data. Alternatively, the data manager is a system that uses at leas one of metadata, attributes, contents, contexts, keywords, histories, heuristics, inference, rules, damarcation, time, date, related handling costs, related handling gains, and sources of the data so as to facilitate the analysis of data, wherein the features is extracted at least partially on the basis of one of a syntactic structure, syntactic attributes, a linguistic structure, and linguistic attributes. The linguistic feature is a system that includes at least one of words generated from a sentence, a word bigram generated from a sentence, and a word trigram generated from a sentence.
Korean Patent Application Laid-Open No. 2002-0028593 (METHOD FOR ISOLATING HARMFUL WORD) discloses a method for isolating harmful word including allowing a computer user to input member information and access a management server, allowing the computer user to input information for registration/transmission in/to an information input screen displayed by the management server, allowing a module for isolating a harmful word of the management server to determine whether a harmful word is included in the information input by the computer user, and allowing the management server to isolate the connection with the computer user or register/transmit the information according to the determination results of the module for isolating a harmful word. Assuming that the harmful word means jargons, slang, ones related to sexual insulation or personality insulation, the Korean Patent Laid-Open No. 2002-0028593 includes a processing module that continuously performs chatting, when the transmitted message does not include the harmful words. The Korean Patent Laid-Open No. 2002-0028593 is described based on whether the message in chatting includes the harmful words.
Korean Patent Application Laid-Open No. 10-2006-0062300 (MULTISTAGE TEXT FILTERING METHOD FOR ISOLATING HARMFUL SITE) discloses a multistage text filtering method for isolating a harmful site that includes dividing texts to be used as data for isolating a lewd site according to a degree of lewdness and the texts as at least one set and then databasing the texts, collecting the texts from an accessing web site, and performing at least one multi-stage filtering by comparing the texts with the texts in the set to determine whether the accessing web site is isolated, thereby isolating the harmful site according to the degree of lewdness within the texts.
Korean Patent Application Laid-Open No. 10-2006-0062300 (A SYSTEM AND METHOD FOR PROVIDING IMPROVED SPAM MESSAGE FILTERING) discloses a method for providing spam message filtering including: a wireless network server that receives a message to be transmitted from a caller terminal to a user terminal, a Corpus DB that stores usage examples of natural language sentences and their affiliated information and is a large scale language DB (database), a vocabulary analyzing unit that detects reception of the message in the wireless network server and extracts at least one noun type keyword from a sentence included in the received message, a Corpus search device that selects any one of the extracted noun type keywords and searches the frequency of spam used in each sentence included in each user category of the Corpus DB and a general frequency used in sentences included in a category other than the user category, when at least one spam frequency and general frequency are inputted, a probability calculation unit that calculates a probability that a sentence including a selected noun type keyword is included in the user category, and a filtering controller that controls the vocabulary analyzing unit and the Corpus search device to make the probability calculation unit calculate the probability that each noun type keyword is included in a spam message, determines whether the received message is a spam message, and transmits the determining result to the radio network server.
Korean Patent Application Laid-Open No. 10-2008-0000416 (FILTERING SYSTEM FOR HARMFUL MESSAGE AND METHOD THEREOF AND RECORDING MEDIUM THEREOF) discloses a technical field for classifying harmful message with a filtering system for filtering harmful message on internet community. Methods thereof are provided to achieve high marketability in an anti-spam market by offering an intelligent spam filtering technology. The filtering system for harmful message comprises a database for storing messages received from clients, a message receiver that receives the messages, a word extracting unit that extracts a plurality of words from the received messages, and an evaluation unit that determines whether the messages are harmful by using the extracted words, stores it in a database, and the evaluation unit uses the word evaluation value stored in the database to determine whether the massage is a harmful message and the message is sorted into a plurality of harmful message.
Korean Patent Registration No. 10-0484944 (SYSTEM FOR AUTOMATICALLY TAGGING SEMANTICS OF MORPHEME BASED ON LOCAL SYNTAX RELATION AND SEMANTIC CO-OCCURRENCE DICTIONARY) discloses an automatic morpheme semantics tagging system based on a local phrase relationship and a semantic co-occurrence dictionary including a short sentence recognizer that recognizes a short sentence by using a basic syntax generation rule, a dependency syntax rule, an electronic morpheme meaning dictionary, a semantics frame and a declinable phrase co-occurrence dictionary generated via a morpheme analysis of an original sentence, an auxiliary word case recovery module that performs a case recovery for an auxiliary word by using a semantics frame and a declinable phrase semantics co-occurrence dictionary, an argument noun meaning determination module that determines a meaning on an argument noun using the declinable phrase semantics co-occurrence dictionary, a noun phrase meaning determination module that determines a meaning on a noun within a noun phrase and a compound noun using a noun phrase meaning co-occurrence dictionary configured of noun/preposition/noun and co-occurrence information of noun/noun. The automatic morpheme semantics tagging system uses the syntactic structure information in view of a linguistic aspect to use the syntax context information that is more accurate co-occurrence information than the existing other semantic ambiguity devices, thereby proposing a method of solving the meaning ambiguity of noun.
Korean Patent Registration No. 10-0757951 (SEARCH METHOD USING MORPHEME ANALYZING IN WEB PAGE) discloses a searching method through morpheme analysis of a web page including displaying the contents of the connected website to the tab browser, If a client user is connected to a website through the tab browser, discriminating a frequency of repeated sentence or words by performing a context morpheme analysis for the contents displayed to the tab browser of a client by an analysis server, selecting the sentence and words having a frequency of a predetermined value or more among the identified words as keywords that are important words, outputting the selected important words as a title of each tab browser, searching the title of each tab browser as a query word of a portal site and outputting the search results as contents of each tab browser, in an analyzing server providing information to the client such as notebook, mobile phone, PDA, or the like. The patent calculates the number of sentences and the number of words that appears in the text with one methodology of the search system by the morpheme processing to output the tab browser title as the keyword.
Korean Patent Registration No. 10-0691400 (METHOD FOR ANALYZING MORPHEME USING ADDITIONAL INFORMATION AND MORPHEME ANALYZER FOR EXECUTING THE METHOD) discloses a method for analyzing morphemes including obtaining morpheme analysis object and additional information associated with the morpheme analysis object from data for index, generating a key based on the additional information, and performing the morpheme analysis on the morpheme analysis object using the key.
Korean Patent Application Laid-Open No. 10-2007-0029389 (METHOD AND SYSTEM FOR PROVIDING ADVERTISEMENT SERVICE USING CORE KEYWORD AND RECORD MEDIA RECORDED PROGRAM FOR REALIZING THE SAME) discloses a recording medium that is typically implemented with command words executable by a digital processing device in order to provide a keyword advertisement service using a core keyword and may be read by a digital processing device, including extracting the keyword by analyzing morphemes of contents text, determining core keywords through interfacing with an advertisement keyword DB server, receiving an advertisement list matching the core keyword through the interfacing with the advertisement DB server, and inserting at least one of the advertisements included in the advertisement list into the contents text.
Korean Patent Application Laid-Open No. 10-2006-0011333 (SYSTEM AND METHOD FOR PROVIDING LOCAL INFORMATION BY MESSAGE ANALYSIS) discloses a system for providing local information including: a database that stores advertisement information of each area and business type, a message recognizing unit that analyzes a short message stored or a text message stored, a voice recognizing unit that analyzes the stored voice message, a location information collecting unit that checks a current location of a subscriber terminal, a service controller that requests whether a message related to a service subscriber has been stored from the message recognizing unit and the voice recognizing unit, checks a received message analysis result, requests current location information of the subscriber terminal from the location information collecting unit, and provides a desired business type and area information according to the message analysis result to the subscriber terminal, and an SMS generation unit that searches information of the database, adds a searched result in a short message, and transmits it to the subscriber terminal.
Korean Patent Application Laid-Open No. 10-2007-0015752 (SYSTEM AND METHOD FOR AD TRIGGERING MESSAGING SERVICE BASED ON MOBILE MESSAGING SERVICE) discloses a system for AD triggering messaging service including a message server transmitting the contents of the message to the receiving terminal after transmitting the predetermined AD message to the caller terminal when detecting the AD triggering identifier by analyzing a message transmitted from a caller terminal.
Korean Patent Registration No. 10-0775680 (METHOD AND SYSTEM FOR PROVIDING ADVERTISEMENT CONTENTS BY USING CHATTING IN MOBILE TERMINAL) discloses a method of transmitting and receiving message through a messenger between a mobile communication terminal and a server through a wireless communication network, including determining whether the mobile communication terminal connected to the server selects any virtual chatting partner of chatting partners, extracting chatting scenario of AD contents good according to the user preference information of the mobile communication terminal in the server to transmit the message according to the scenario rule to the mobile communication terminal, transmitting a response message for the message of the server in the mobile communication terminal, analyzing the message of the mobile communication terminal received in the server and transmitting the message requesting the AD contents information recommendation permission according to the user's intention and the scenario rule to the mobile communication terminal, transmitting the AD contents information to the mobile communication terminal when transmitting the message permitting the AD contents information recommendation from the mobile communication terminal in the server, and adjusting the character state of the messenger to correspond to the user's intention by analyzing the message of the mobile communication terminal in the server prior to the transmitting of the message to the mobile communication terminal.
Korean Patent Registration No. 10-0597435 (SYSTEM AND METHOD FOR CLASSIFYING QUESTION BASED ON HYBRID OF INFORMATION SEARCH AND QUESTION ANSWER SYSTEM) discloses a system and method for classifying a question based on a hybrid of an information search and question answer system including a questioned work name processor that recognizes the work name included in the inputted question, a question language analyzer that converts each word into a semantic code through a morpheme analysis process, an object name recognition process, and a word meaning tagging process, a rule-based question classifier that classifies the answer type requested from the question by using an LSP (Lexico-Semantic Pattern) code of the question and a predefined question classification rule, a statistics-based question classifier that classifies the LSP code of the question by using statistics information constructed in a learning document tagging the answer type, and a questioned answer type determiner that finally determines the answer type of the user question by using a result of both classifiers.
Korean Patent Registration No. 10-0361166 (SYSTEM AND METHOD FOR SEARCHING INFORMATION) discloses a system and method for searching information including a processing device that codes input words and searches the data base based on the coded words to search the information corresponding to the code words when inputting words indicating information to be searched or a field to which the corresponding information belongs after a data base that sorts a plurality of information according to field, allocates codes to each sorted field, and codes words indicating a plurality of information included in the corresponding field in association with a code of each field accesses a user searching data via a network using a communication device, wherein the information stored in the database is sorted into the plurality of information regions, the information regions are each allocated with an information sorting code which is sorted into at least one detailed information region, the detailed information regions are coded in association with the information sorting code of the corresponding information regions, the words having meanings equal or similar to each other among the words indicating the corresponding information are stored by the same code, each word code is arranged in a set order.
Korean Patent Application Laid-Open No. 10-2005-0092955 (SYSTEM AND METHOD FOR ONLINE AD INCLUDING COPY OF ADVERTISER FIT FOR CONTENTS) discloses a system and method for advertising online including a first storage device in which information to be advertised is stored; a second storage device in which original contents are stored; at least one server that fetches the original contents from the second storage device, analyzes the original contents by a natural language processing mechanism, and fetches the information to be advertised corresponding to the analysis results from the first storage device and inserts a portion of the information among the advertisement information into the original contents, a third storage device in which the contents inserted with the advertisement information is stored by the at least one server.
Korean Patent Registration No. 10-0669534 (METHOD AND SYSTEM FOR EXTRACTING SUMMARY OF DOCUMENT BY UTILIZING SENTENCE ABSTRACTION AND ABDUCTION RULES) discloses a method of summarizing a document using sentence abstraction and abduction rules: including sentence inputting that inputs and stores document contents to be summarized, syntax analyzing that analyzes and stores a syntax for data stored in the step; sentence main component extracting that sequentially reads the syntax data at the step and stores ontology data; sentence abstracting that understands each of the selected main components as an abstraction meaning by comparing the ontology data value at the step with a recognition upper range information that is a predetermined sorting reference for meanings included in each of the selected main components; topic sentence selecting that reads the stored value at the step and input the read stored value to an abduction rules program movable device unit to select a plurality of topic sentences having abduction between syntaxes and store each identification value for the plurality of selected topic sentences; summary writing that reads corresponding values of abstraction meaning understood for the stored and selected main components at the step and combines and stores the read values as a grammar summary sentence; a summary sentence outputting that determines whether there is the output control signal and if so, outputs an displays the data value stored at the step through the operation of the output unit.
Korean Patent Registration No. 10-0836878 (APPARATUS AND METHOD FOR ALLOCATION OF SUBJECT OR FIELD IN INFORMATION SEARCH SYSTEM) is an apparatus for allocating a subject or a field in an information search system that is connected to a user terminal, a web portal site, a web site, searches information, and provides the searched information, the information search server including a search engine that performs information search for a document based on an index word corresponding to a query word or a document; a database that stores and manages information searched in the search engine; a thesaurus matching unit that extracts the index words from an original sentence stored in the database and performs thesaurus matching by the index word to perform the subject or filed allocation for the original sentence; a folksonomy processor that extracts terminologies from a definition sentence stored in the database and allocates the subject by the extracted terminologies and uses folksonomy to allocate a field.
In addition, the information search system used by a large information search portal such as Google, Yahoo, or the like, is based on a search key type of a “word” unit. Although a ‘phrase’ input may be performed, a current technology connects an N-gram based index key by a logical operator (AND, OR, NOT) rather than a language processing method according to a high technology of the “phrase” to propose the search results, such that there is a limitation in improving the reproduction and accuracy of the information search. In particular, a ‘mobile information search’ technology has been in the limelight as a next generation information search field and a new web search technology is being proposed and discussed as a “semantic web’ technology. However, there is no special successful example up to now.